I

R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

Parametric Investigation of Routing Protocols in Wireless

Sensor Networks in Different Scenarios

SAJJAD ABBAS
MS Scholar; Department of Computer Science
COMSATS Institute of Information Technology Islamabad, Pakistan
E-mail: [email protected]
Tel: +92- 03457442432

ISRAR AHMAD
MS Scholar; Department of Computer Science
Virtual University Islamabad, Pakistan
E-mail: [email protected]

MUHAMMAD AFZAL BHATTI

MS Scholar; Department of Computer Science
Islamia University Bahawalpur, Pakistan
E-mail: [email protected]

NAEEM AKHTAR
Lecturer; Department of Management Sciences
University of Education Lahore (Okara Campus), Pakistan
E-mail: [email protected]

Abstract
Modern developments in the field of micro-sensor devices have enhanced research and developments in the
sensor networks leading to many new protocols are being designed especially for Wireless Sensor
Networks (WSNs). Wireless sensor networks with hundreds to thousands of sensor nodes can collect data
from different location and transmit the collected data to a particular user, depending on the application
and its requirements. These sensor nodes have some constraints due to their limited energy, storage
capacity and computing power. Information is routed by various routing protocols from one device to
another device. Wireless Sensor Network is a challenging network in which communication between nodes
is unpredictable and difficult. This demands the performance analysis of different routing protocols to find
out a protocol which more suitable for a particular scenario. This Research will present the performance
evaluation of reactive (AODV, DSR) and proactive (DSDV) routing protocols in WSNs by varying
parameters like node density, node speed and pause time for TCP and UDP traffic patterns. Under
different network scenarios and Mobility models, these routing protocols will compare using widely used
performance metrics such as packet loss, end-to-end delay and throughput.

Key Words: Wireless Sensor Networks (WSN), Mobile Ad-hoc Networks (MANET), Vehicular Ad-hoc
Networks (VANET), Topology, Proactive, Reactive, Hybrid.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 49

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

Introduction
Sensor technology is very important in today's growing industrial and life in rats. Automatic and semi-
automatic equipment to adjust and control the number of process depends on the sensors. Furthermore,
safety system, sensor environmental conditions and to monitor the structural integrity of the building part is
very significant. Conventionally, the sensor signals from these sensors are processed, and must be
interpreted as part of a computing device connected to the central conductor through some wires. Severe
sensor wiring is expensive, limiting the flexibility and reusability of the sponsors (Chowdhury, 2009).

Progress of wireless communication and data processing microchip sensing techniques may be combined to
produce a compact device is now possible. This device is actually called Mott and it is miniature device. As
they pass through the radio; link is unnecessary but no need of wires to transfer the files. Mott filter
processing data exchange itself and/or it may be combining it. Wireless communication works within a
limited range. Until it reaches the destination device, effectively extending the range of all motes; rules
based in part on the multi-hop routing protocol message can be passed (Halpern, 2002). This remote area of
dozens, hundreds, thousands or wireless sensor network (WSN) is created to transfer the data among all
connected devices.

In WSNs, in order to monitor temperature, pressure, vibration and other environments, such as the health of
pollutants or sound, collectively, such as motion sensor nodes arranged in a separate work area with a lot of
wireless sensors nodes . The biggest reason is the granite on the battlefield at the beginning of the
development of military applications of WSNs but now days its applications are being expanded to control
field, status monitoring traffic, and monitor industry. Because of the bandwidth depends on the memory
and the operation speed of the sensor nodes (Balakrishnan, 2002), the results of the other constraints of the
energy cost and size limitations are very crucial.

Wireless sensor nodes in any network are composed of the following components:

 Microcontroller
 Radio transceiver
 Energy (Battery)

WSNs, a number of unique features, such as the following:

 Scalability.
 Deployment in large geographical area.
 WSNs can handle node failures.
 Another feature of the mobile node that there is a variety of environments, and their ability to
survive in the environment.
 They have a dynamic network topology.

1.2 Wireless Communication

Growth in the field of wireless commutation is becoming very popular now days. Due to fact that Wireless
technology is applicable in different situations; wireless nodes has the ability to reach in any area of the
environment. Wireless devices are now available in a vast variety of market. A lot of wireless networks are
used to communication among different nodes; this network may be infrastructure network or
infrastructure-less network. MANET, VANET, cellular network are some examples of wireless network
(Heidemann, 2003).

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 50

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

1.3 Wireless Communication Requirements

Wireless communication can be done using some wireless protocols; these protocols are being built up
from the raw electro-magnetic singling (Mukherjee, 2011). Radio Frequency (RF) carrier is modulated
cautiously using some transmitter. Receiver is used to demodulate the signals as well as demodulator
performs analysis of the signals.

Fig. 1 demonstrates main characteristic and main phases of the packet switching network using wireless
communication protocol. Operations in wireless communication are done by the wireless communication
protocols. In wireless communication can be seen using different layers of the wireless communication
phases.

Wireless Communication Phases

Transmission

Data to be Transmitted

Encode processing

Start Symbol
Transmission Encoded data to be Transmitted

MAC Delay Transmitting encoded bits

Bit Modulation

Radio Samples

Start Symbol Receiving individual bits

Search

Start Symbol Detection Synchronization

Encoded data received

Decode processing

Data Received

Fig. 1: Phases of wireless communication for transmission and reception

1.4 Wireless Sensor Node Architecture

Sensor nodes are low-cost, low-energy conservation, multi-functional nodes which have small form factor
and can communicate unstrapped using radio frequency communication over a short distance (Kamforoush,
2009). WSNs are deployed to monitor the physical environment and network nodes can be deployed
anywhere.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 51

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

Fig. 2: Architecture of a Sensor Node

1.4.1 Architecture and constraints

Wireless Sensor Network (WSN) consists of the following components (Bikas, 2009):

The Direction Detection:

Depending on the sensor and the sensor unit is similar in general-to digital converter (ADC). However,
unlike the detection unit is incorporated in the sensor node, it is not necessarily true.
Processing Unit:

In general, the processing unit is associated with a small storage parts and processes sensor data reading, in
cooperation with other sensor nodes.

Transceiver Unit:

The transceiver sensor node with other nodes that are connected Network - for example, using a vertical
cavity, a radio frequency, infrared (IR), optical media, Use optical communication by the surface emitting
laser or VCSEL (VCSEL), and radio.

Power Unit:

A major power, it follows the power unit is a very important component sensor is a sensor network Node.
A solar cells such carriers, and is provided by the energy recovery unit, or Use of the chemical battery.

Constraints:

While we talk about WSN, there are some constraints and they playing specific role assigned to them (Kato
, 2011):

 All the elements above must be part of a small unit.

 Low power consumption, significant other constraints, high-capacity "Network density ", and
fault-tolerant.

1.5 Problem Description

Performance plays a vital role in any network to provide its best services to its users to get maximum
satisfaction level. Performance measurements are a difficult task in real time environment. While using
any network services; network simulators are available to evaluate the network performance for the service
provider as well as for its users. This research paper investigates and evaluate WSN performance by
ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 52
I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

selecting parameters like delay, throughput, data drop that are evaluated in NS2 with scalability in the
network by increasing its size and then a comparison between the two is made to determine which protocol
works best in the required application.

The protocols which are chosen to evaluate the performance in WSN are AODV, DSR, and DSDV (Kato,
2007). The reason why WSN were chosen for this research is because the configuration of ad hoc has
greater potential in use that is widespread among the wireless network. This includes low cost, provision of
higher bandwidth and it is also scalable compared to cellular communication.

1.6 Proposed Solution/Methodology

Main aim of this paper is to explore the major issues in WSN particularly network delay, data drop and
throughput. Literature review is done thoroughly to get deep understanding about the above said issues. To
enhance more knowledge diffident International journals, conferences, online resources, different
discussion forums and blogs are consulted. In this study parametric investigation is done by well-known
and authenticated network simulator NS2. Analysis of the results has done to prove the problem statement.
Moreover, performance is evaluated using different mobility models. I have used Manhattan Grid Mobility
Model which is one of the most widely used Group Based Mobility Model. In Manhattan Grid Mobility
Model nodes are divided into two groups with predefined paths.

2. Routing Protocols

2.1 Wireless Sensor Routing Protocols

The WIRELESS SENSOR NETWORK protocols are divided into five categories namely:

 Position based
 Topology based
 Broadcast routing protocol
 Cluster based
 Geo cast routing protocol

It is based on the application and regions, their characteristics are most suitable for them. Unlike
WIRELESS SENSOR NETWORK supports a place, which is unlimited, high WIRELESS SENSOR
NETWORK nodes are the random motion of the mobile node, some traffic, continue to allow you to follow
the pattern to its movement (Dash & Chlmtac, 2004). Despite the changes in the topology less imitation,
the second of the leading drivers, the rest must be based on their driving behaviour and topology of
different speeds are dynamic, highly complex ways.

Because of significant differences between the ad hoc networks, the study must be interrupted in order to
confirm compatibility reasons WIRELESS SENSOR NETWORK environment for ad hoc network routing
protocol is completed. Routing protocols are enrolled in the study from the so-called WIRELESS SENSOR
NETWORK routing protocols, topologies based on conventional protocol specific sectors. The main
reason, it must make a choice is a dynamic topology WIRELESS SENSOR NETWORK embodiment has a
direct impact on the routing protocol (Mittal, 2013).

2.2 Topology Based Routing Protocols

Its main purpose is in information linkages in existing network data sending (Farman, 2012).

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 53

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

Topology
based
Routing
On-
TableProtocols
Deman Hybrid
Driven
d

CGS ABR ZRP

DRF AOD ADV

WRP ARP

DSD BSR

DTD DSR

HSL FOR

HSR LBR

LCA LMR

OLS SSR

STA SMR

TOR

PLB

Fig. 3: Topology based Routing

Protocols
2.2.1 Types of Routing Protocols:

 Reactive (on – demand) routing protocol

 Proactive (table – driven) routing protocols

2.2.2 Proactive or Table Driven Routing Protocols

Based on the shortest path, consensus algorithm mainly positive. Further, shared by all nodes connected in
a table form contains information between their neighbours and Track has not been used, they maintain,
and provides a vehicle type all, in spite of not being used by the updated node continuously Road transport
between them, the routing updates on the network I was given several times. Never seen in making the size
of the network bandwidth limitations routing updates periodically, and the network load (Saket, 2011).

2.2.3 On Demand or Reactive Routing Protocol

These protocols are used to solve the routing protocols established by the initiative of overhead problem.
This is achieved by maintaining a positive way completed. On the basis of need, these protocols, the
demand for the use of a specific path. They also maintain the only road currently being used only when the
load on the road network which will reduce the available subset at any time. Type in some of the agreement
is that these TORA AODV, DSR, JARR and PGB (Singla, 2011).

2.2.4 Proposed WIRELESS SENSOR NETWORK Topology Based Routing Protocols

Details about these protocols given below:

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 54

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

2.2.5 Ad-hoc On Demand Vector (AODV)

AODV protocol improved DSDV protocol. If two nodes need a connection node routing table and to keep
the whole, the only way to establish the AODV protocol. Classification is the on-demand routing protocol
AODV (Saket, 2011) acquisition system as pure. This is because it is stored in the node is not in the path, it
is selected a few routing information.

Problem using serial numbers route discovery process to avoid AODV "unlimited count" function. AODV
routing messages used in the process, in response to the root routing error (RERR), and a route request
(RREQ) 3 kinds of following the (RREP). Network communication needs to remain silent communication
node by the source node. It was found, when a valid path does not exist, start the route discovery process
begins. Then, until you reach a point of intermediate nodes or the RREQ purpose (Singla, 2011), sent to
their neighbours RREQ is a valid route to the destination, the source node to continue when the RREQ
forwarded to the neighbours. Broadcast ID and serial number assigned to each node. Increase the sequence
number of each generation is generated along the path of the RREQ. Therefore, the RREQ identified, based
on the source node, and based on the broadcast ID and IP address (Rao, 2010). It returns RREQ revoked,
when it reaches the intermediate node or the route to the destination is active. From another node, RREQ
all other later arrival is outdated. RERR link failure detection is not started after being fed back to the
source node; beginning with the discovery of a new route from the network receives a RERR upstream
node process.

2.2.6 Dynamic Source Routing:

Based on the concept of source routing, DSR is a simple request to run the ad hoc routing protocol
agreement on it. It is based on the routing information for each node, but the cache because it requires
properties. The presence of a routing network they have been updated, cached routing information
indicating a route search Yes. Because the routing information periodically transmit information cache is
overhead is minimized. Two processes, route maintenance and route discovery is a DSR protocol mainly
(Joon, 2012).

I route the RREQ messages used by AODV protocol, the discovery process and DSR RREP. Before
sending the remaining RREQ, to determine if there may be some ways, the source node asks its path cache.
In the presence of a shortage of route cache that the source node sends RREQ. This ID number is included,
is a unique address for the destination and source. RREQ When you get it is determined by the node with a
valid route to the destination. Arriving in the way destination node and the final destination, the RREQ, it's
on to the next node through sequential steps. The RREQ part, the ID number of the same with, but as a
mechanism for receiving the broadcast source node difference in the way the record by the nodes on the
path. What about the RREQ forwarding nodes are not written in the RREQ destination address (Man, 2001)
only his was situation. When keeping the route has not expired path, it reaches the intermediate node to the
first, and then have to produce, RREP transmission based on the routing information. For intermediate
nodes, the routing information, it will be added RREP.

2.2.7 Destination-Sequenced Distance Vector

Bellman-Ford routing algorithm is based on the classic table-driven routing protocol DSDV. Serial number
for the Bellman-Ford algorithm, which includes improvements in a routing table by the free loop. If the
number of hops in the network and the destination of those memory locations as possible, where each of all
the mobile nodes in the system have a routing table [16]. Each path / paths to the path marked with a serial
number and the highest sequence number used in the case where the destination. It is helpful to identify the
new state of the route you through this avoids the formation of loops. "Full tilt", "increment" is a system
that minimizes the traffic generated, is a "full dump is the implementation of the relevant information

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 55

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

packet" changes. Two types of packets do, only occasionally "increase" is used which leads to increased
transmission efficiency of the overall system during training.

 Data broadcast, the following is to include all the nodes in the mobile.
 New serial number
 Previous address.
 Number of hops to reach the destination.
 About the destination of the received sequence number information.

If you want to increment the serial number of each node's message. Find out whether the instance node,
node A cannot reach the target sector, terminate in the neighbouring odd cause infinite metric number. If a
node that receives a digital endless measuring and updates its table to ensure that the routing is not
infinitely high latency sequence number comparison metric counting until it receives. Everything has to be
calculated for each mobile host as the weighted average of the optimal route.

2.3 Topology Based Routing Protocols Comparison

Table 1: WIRELESS SENSOR NETWORK Topology Based Routing Protocol Comparisons

Protocol Property DSDV DSR AODV
QoS Support No No No
Loop Free Yes Yes Yes
Multicast Routes No Yes No
Distributed Yes Yes Yes
Unidirectional Link Support No Yes Yes
Multicast No No Yes
Periodic Broadcast Yes No Yes
Routes maintained in Route Table Route Cache Route Table
Route Cache/Table Timer Yes No Yes
Reactive No Yes No

3. Simulation Results
Simulation Parameters
Table 2: Simulation Parameters
Parameter Value

Protocols AODV,DSDV,DSR
Simulation Time 500 s
Number of Nodes 10, 15, 20, 25, 30, 35, 40
Simulation Area 500 m x 500 m
Pause Time 0, 50, 100, 150, 200, 250, 300, 350, 400 s
Connection Type TCP,UDP
Maximum Speed 10, 20, 30, 40, 50 m/s
Mobility Model Random Waypoint, Manhattan Grid
Network Simulator NS 2.35

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 56

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

3.1 Ad-hoc On-demand Distance Vector (AODV)

AODV routing protocol is a reactive routing protocol which was first proposed by an IETF Internet draft in
1997. AODV is an improvement of DSDV and is collectively based on DSR and DSDV. As indicated by
the name of AODV it works on demand basis i.e. it does not maintain route information without need.

Operation of AODV:

Operation of AODV done by the following two methods:

1. Route Discovery:

This is a process in which a source node wishing to send a packet to a destination whose address, if not
present in the routing table of the source node (sender) then the source node dynamically obtains the
address by the mechanism of Route Request (RREQ) and Route Reply (RREP).

2. Route Maintenance:

When the route is established, the source node will maintain it as long as it is needed.

3.2 AODV Traffic Analysis with respect to Node Density:

In this first section the impact of node density on the performance of TCP and UDP over AODV in term of
Throughput, Delay and Packet Loss is discussed.

AODV Throughput Analysis with respect to Node Density: Fig. 4 shows two line graphs which depicts
the throughput of TCP and UDP over AODV. The results show that the throughput of TCP and UDP
increases by increasing the nodes. Throughput is directly proportional to node density. Individually, the
throughput of TCP is higher than that of UDP. The reason behind the high throughput of TCP is default
size of TCP window and given data rate (0.1 Mbps).

TCP Throughput
UDP Throughput
700
600
Throughput(kbps)

500
400
300
200
100
0
10 15 20 25 30 35 40
Node Density
Fig. 4: Throughput Comparison of AODV over TCP & UDP Connection in case of Node density

AODV Delay Analysis with respect to Node Density: The given line graphs in Fig. 5 show that node
density has very little impact on the delay of the AODV over TCP. From node 10 to 25 the delay is
relatively same and beyond the node 25 the delay is slightly decreased. Node density has a very fine effect

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 57

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

on the delay of AODV over UDP. Delay of UDP has inversely proportional relationship with node density
i.e. by increasing the nodes the delay decreased.

TCP Delay
UDP Delay

7
6
5
Delay(sec)

4
3
2
1
0
20 25 30 3510 15
40
Node Density
Fig. 5: Delay Comparison of AODV over TCP & UDP Connection in case of Node density

AODV Packet Loss Analysis with respect to Node Density: The simulation setup shows in line graphs at
Fig. 6 that node density has no certain effect on packet loss of TCP. The packet loss ratio of TCP is about
same from node 10 to node 40. The reason is that TCP retransmit packet every time and hence no certain
loss occur in TCP as compared to UDP. UDP has highest packet drop ratio than TCP. By increasing the
number of nodes the packet loss of UDP become decreased. The reason UDP highest drop rate is that the
sender sends data continuously without knowing the receiver buffer and also there is no acknowledgement
from the receiver. In this way the receiver can’t accommodate further packets and finally lost it.

TCP Packet loss

UDP Packet loss
40
Packet loss(nos of packets)

35
30
25
20
15
10
5
0
10 15 20 25 30 35 40
Node Density

Fig. 6: Packet loss Comparison of AODV over TCP & UDP Connection in case of Node density

3.3 AODV Traffic Analysis with respect to Node Speed:

Node speed is the sophisticated property of nodes in MANET. Here I discuss the effect of node speed in
terms of Throughput, Delay, and Packet Loss.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 58

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

AODV Throughput Analysis with respect to Node Speed: Fig. 7 shows that overall TCP and UDP both
are highly affected by the speed of nodes. The throughput performance of TCP and UDP both degraded by
increasing speed of the nodes. TCP throughput at 20 m/s is 562.8 Kbps and on 40 m/s the throughput is 561
Kbps means the performance is being degraded. Similarly the throughput of UDP at 20 m/s is 547.6 and at
40 m/s the throughput is 502.2 Kbps. Individually, TCP has higher throughput as compared to UDP.

TCP Throughput
UDP Throughput
580
560
Throghput(Kbps)

540
520
500
480
460
440
420
10 20 30 40 50

Node Speed
Fig. 7: Throughput Comparison of AODV over TCP & UDP Connection in case of Node Speed

AODV Delay Analysis with respect to Node Speed: Fig. 8 depicts that node speed has no too high effect
on the performance of TCP as we see that from 10 m/s to 50 m/s the result is round about same. While in
case of UDP, the performance in low speed is better than higher speed. The delay in 10 and 20 m/s is less
than beyond the 20 m/s.

TCP Delay
UDP Delay

6
5
Delay(Sec)

4
3
2
1
0
10 20 30 40 50
Node Speed
Fig. 8: Delay Comparison of AODV over TCP & UDP Connection in case of Node Speed

AODV Packet Loss Analysis with respect to Node Speed: Fig. 9 shows that UDP Packet Loss rate is
very high than the TCP. This has same reason which I have discussed that in UDP the lost packets are not
retransmitted and also packets are send to receiver without knowing receiver receiving rate. TCP has very

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 59

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

lowest packet lost ratio because unlike UDP, TCP set window size according to the receiving rate of the
receiver.

TCP Pkt loss

UDP Pkt loss
40

Packet loss(Nos of Packets)

35
30
25
20
15
10
5
0
30 40 10
50 20
Node Speed
Fig. 9: Packet loss Comparison of AODV over TCP & UDP Connection in case of Node Speed

3.4 AODV Traffic Analysis with respect to Pause Time:

Throughput, Delay and Packet Loss is evaluated with varying pause time.

AODV Throughput Analysis with respect to Pause Time: Fig. 10 shows that there is no big difference
between the TCP and UDP throughput. On multiple points the TCP and UDP throughput is nearest to one
another. However the cumulative throughput of TCP is higher than UDP. By increasing pause time the
throughput of AODV over TCP and UDP both is degraded.

TCP throughput
UDP Throughput
700
600
Throughput(Kbps)

500
400
300
200
100
0
0 50 100 150 200 250 300 350 400
Pause Time
Fig. 10: Throughput Comparison of AODV over TCP & UDP Connection in case of Pause Time

AODV Delay Analysis with respect to Pause Time: Clearly the Fig. 11 depicts that UDP has highest
delay as compared to TCP. Individually, there is no big effect of Pause Time on the performance (delay) of
AODV over TCP as delay from 0 to 400 second is very minor. While UDP performance degraded with
increased number of Pause Time i.e. UDP delay increases by increases the Pause Time.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 60

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP Delay
UDP Delay
10
8

Delay(sec)
6
4
2
0
0
50 100 150 200 250 300 350 400
Pause Time
Fig. 11: Delay Comparison of AODV over TCP & UDP Connection in case of Pause Time

AODV Packet Loss Analysis with respect to Pause Time: Fig. 12 shows that AODV over TCP is less
affected by the Pause Time while UDP is highly affected from the Pause Time. The Packet loss ratio of
UDP is higher than the TCP due to the reason discussed above. AODV over UDP packet loss ratio is
increased by increased number of pause time.

TCP Pkt loss

50
Packet loss(Nos of

40
Packets)

30
20
10
0
50 100 150 0
200Time
Pause 250 300 350 400
Fig. 12: Packet loss Comparison of AODV over TCP & UDP Connection in case of Pause Time

3.5 Destination Sequence Distance Vector (DSDV)

Destination-Sequence Distance-Vector is a proactive (table driven) routing protocols in MANETs. It was

developed by C Perkins in 1994. This protocol works on the basis of Bellman Ford algorithm. In DSDV
network, all the nodes (computers) have the entire topology of the network i.e. each node knows about its
neighbours, next hop, shortest path, final destination etc. The nodes contain all these information in their
routing table by the regular updates. By regular updates, the packets can be easily sent to the destination
because of all known path but regular updates is one the cons of the DSDV as it consume much of the
energy of the network because of the regular updates. The DSDV use sequence number with the help of
which it avoid loops and separate stales routes from new routes.

3.6 DSDV Traffic Analysis with respect to Node Density:

Node Density is the number nodes participating in network (MANET). I analyze the status of throughput,
delay and packet loss using node density.

DSDV Throughput Analysis with respect to Node Density: In Fig. 13 the simulation results show that
the throughput of network both in TCP and UDP is directly proportional to the node density. Increasing
number of nodes increases the throughput of DSDV over TCP and UDP both. Overall, the throughput of
ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 61
I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP is higher than the UDP. This is because the given data rate is small in size as compared to the size of
the window of TCP. That’s why the data can pass via the TCP window with no chance of losing the
packets.
TCP Throughput
UDP Throughput
700

Throughput(kbps)
600
500
400
300
200
100
0
20 25 30 35 40
Node Density 10 15
Fig. 13: Throughput Comparison of DSDV over TCP & UDP Connection in case of Node density

DSDV Delay Analysis with respect to Node Density: From the simulation results, Fig. 14 shows that
TCP delay is independent of node density while UDP is being affected by node density. Increasing number
of nodes has very minute impact on delay of the DSDV over TCP while in case of UDP; more nodes have
more delay as compared to less number of nodes.

TCP Delay
7
6
Delay(sec)

5
4
3
2
1
0
20 Density
Node 25 30 1035 4015
Fig. 14: Delay Comparison of DSDV over TCP & UDP Connection in case of Node density

DSDV Packet Loss Analysis with respect to Node Density: Clearly Fig. 15 depicts that the packet loss of
UDP is very higher than TCP. Individually, the effect of node density on TCP packet loss is minor or
sometimes very negligible. This is because TCP by nature, retransmit packet every time and that’s why it
has very smallest packet loss ratio. On other hand, UDP has highest packet loss because of sending
continuous packets without knowing receiver receiving buffer and no acknowledgement from receiver.
TCP Packet loss
UDP Packet loss
50
Packet loss(nos of

40
packets)

30
20
10
0
20
Node 25 10
30 35 40
Density 15
Fig. 15: Packet loss Comparison of DSDV over TCP & UDP Connection in case of Node density

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 62

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

3.7 DSDV Traffic Analysis with respect to Node Speed:

I evaluated the node speed for analysing the performance of the same metrics i.e. throughput, delay, packet
loss.

DSDV Throughput Analysis with respect to Node Speed: In Fig. 16, line graphs show high throughput
for TCP and low for UDP. By increasing node speed TCP has almost no change while UDP decreased by
increasing node speed. Hence from the simulation setup I concluded that in highly mobility environment,
TCP should be used for high throughput and UDP should not be used because of high packet loss.

TCP Throughput
UDP Throughput
700
600
Throghput(Kbps)

500
400
300
200
100
0
20 30 4010 50
Node Speed
Fig. 16: Throughput Comparison of DSDV over TCP & UDP Connection in case of Node Speed

DSDV Delay Analysis with respect to Node Speed: Fig. 17 shows that for high speed of nodes
performance degraded by UDP i.e. by increasing speed delay in UDP increases while increasing nodes does
not affect the performance of TCP. So I clearly observed and concluded that for high speed nodes TCP
should be used.

TCP Delay
UDP Delay
7
6
5
Delay(Sec)

4
3
2
1
0
30 1040 20
50
Node Speed
Fig. 17: Delay Comparison of DSDV over TCP & UDP Connection in case of Node Speed

DSDV Packet Loss Analysis with respect to Node Speed: Packet loss of DSDV over UDP is maximum
than DSDV over TCP. UDP has highest packet loss ratio because in highly mobility environment the node
has high chances to leave the range. Similarly UDP don’t have the concept of packet retransmission also in
UDP the sender don’t take care of receiver receiving buffer. Hence much lose in packets occurs.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 63

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP Pkt loss

UDP Pkt loss
45
40

Packet loss(Nos of Packets)

35
30
25
20
15
10
5
0
10 20 30 40 50
Node Speed
Fig. 18: Packet loss Comparison of DSDV over TCP & UDP Connection in case of Node Speed

3.8 DSDV Traffic Analysis with respect to Pause Time:

Pause Time is also mostly used factor in MANET. It refers to the stopping time of a mobile node for
communication. I evaluate the same metrics on varying pause time.

DSDV Throughput Analysis with respect to Pause Time: It has been shown in Fig. 19 that the
throughput of TCP is higher than UDP. The reason is discussed earlier. On some pause time specifically
350 and 400 UDP has crossed the TCP throughput.

TCP throughput
UDP Throughput
700
600
Throughput(Kbps)

500
400
300
200
100
0
0
50 100 150 200 250 300 350 400
Pause Time
Fig. 19: Throughput Comparison of OLSR over TCP & UDP Connection in case of Pause Time

DSDV Delay Analysis with respect to Pause Time: Pause time has a very little impact on DSDV over
TCP. This is shown in fig. 20. On other hand pause time has high effect on DSDV over UDP. By
increasing the pause time, the delay in DSDV over UDP increases accordingly.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 64

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP Delay
UDP Delay
20

15

Delay(sec)
10

0
50 100 150 200 250 300 350 4000
Pause Time
Fig. 20: Delay Comparison of DSDV over TCP & UDP Connection in case of Pause Time

DSDV Packet Loss Analysis with respect to Pause Time: The simulation results show that pause time has
very minor impact on DSDV over TCP while DSDV over UDP is highly affected by pause time. This is
shown in Fig. 21 Packet loss of UDP at lower value of pause time is higher than the lower value of pause
time.

TCP Pkt loss

UDP Pkt loss
45
Packet loss(Nos of Packets)

40
35
30
25
20
15
10
5
0
0
50 100 150 200 250 300 350 400
Pause Time
Fig. 21: Packet loss Comparison of DSDV over TCP & UDP Connection in case of Pause Time

3.9 Dynamic Source Routing (DSR)

Dynamic Source Routing (DSR) is a reactive routing protocol which is based on source routing. It does not
use periodic updates like AODV hence it minimize routing overhead. It is specifically developed for multi-
hop network of mobile nodes. DSR saves the bandwidth and conserve the battery life. In DSR, the packet
being sent from source to destination contains the header which consists of all the addresses of the
intermediate nodes between the source and destination.

Packet from S to R
Sender Receiver
A, X, Y, Z, B
S
Fig. 22
R
In the above figure, the sender A sends a packet to receiver B. The packet (represented by the rectangle)
contains the sender, receiver and all the intermediate nodes addresses which are X, Y, Z lies between A and
B.
ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 65
I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

This is the example of source routing i.e. the source initiate the whole header in the packet and sent it to the
receiver.

3.10 DSR Traffic Analysis with respect to Node Density:

With the help of varying node density I find the Throughput, Delay and Packet Loss of the network.

DSR Throughput Analysis with respect to Node Density: Fig. 23 shows that Node Density effect both
TCP and UDP over DSR. By increasing nodes, throughput of TCP and UDP both increased. Individually
the throughput of TCP is higher than the throughput of UDP. The reason of high throughput of TCP is the
default window size of TCP and given data rate (0.1 Mbps). UDP has lower throughput than TCP as UDP
has no proper flow mechanism (no acknowledgement and no windowing concepts) and hence due to much
packet loss it has lower throughput.

TCP Throughput
UDP Throughput
700
600
Throughput(kbps)

500
400
300
200
100
0
10 15 20 25 30 35 40
Node Density
Fig. 23: Throughput Comparison of DSR over TCP & UDP Connection in case of Node density

DSR Delay Analysis with respect to Node Density: Fig. 24 clearly depicts that the effect of Node Density
on the performance of TCP over DSR is minute while the effect of Node Density on the performance of
UDP over DSR is more as compared to TCP. UDP delay increases by increasing nodes. By increasing
nodes, UDP performance becomes degraded. From the simulation results I concluded that in high denser
network TCP should be used for best performance (lowest delay).

TCP Delay
UDP Delay
14
12
10
Delay(sec)

8
6
4
2
0
20 2510 30 15
35 40
Node Density
Fig. 24: Delay Comparison of DSR over TCP & UDP Connection in case of Node density

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 66

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

DSR Packet Loss Analysis with respect to Node Density: The line graphs from Fig. 25 depicts that Node
Density don’t have certain effect on TCP performance in case of Packet Loss. From Node 10 to Node 40
the result is same and approximately equal to zero. Figure 3 also shows that increasing nodes does not
increase packet loss ratio for UDP. Overall, the packet loss ratio of UDP is much higher than the TCP.
TCP Packet loss
UDP Packet loss

60
Packet loss(nos of packets) 50
40
30
20
10
0
20 25 30 35 40 10 15
Node Density
Fig. 25: Packet loss Comparison of AODV over TCP & UDP Connection in case of Node density

3.11 DSR Traffic Analysis with respect to Node Speed:

Here I evaluate Throughput, Delay and Packet Loss in term of Node speed.

DSR Throughput Analysis with respect to Node Speed: Fig. 26 shows Node the comparison of
Throughput Vs Node Speed. The figure reveals that TCP and UDP both are affected by the mobility of
nodes. According to the simulation results, we see that the effect of mobility on TCP performance is minor
than UDP as on 10, 20 and 50 m/s the throughput is 551, 567 and 564 Kbps respectively whereas at 30 and
40 m/s the throughput is little lower than these. UDP throughput is more in less speed and less throughput
in more speed.

TCP Throughput
UDP Throughput

600
Throghput(Kbps)

500
400
300
200
100
0
30 40 50 10 20
Node Speed
Fig. 26: Throughput Comparison of DSR over TCP & UDP Connection in case of Node Speed

DSR Delay Analysis with respect to Node Speed: Fig. 27 show that UDP has more degraded
performance than TCP and also from the figure we see that UDP delay after 30 m/s increases. In case of
TCP there is very smallest difference among the values from 10 to 50 m/s. So according to the conducted
experiments I concluded that Node Speed has no certain effect on the performance of TCP.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 67

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP Delay
UDP Delay

12
10

Delay(Sec) 8
6
4
2
0
10 20 30 40 50
Node Speed
Fig. 27: Delay Comparison of AODV over TCP & UDP Connection in case of Node Speed

DSR Packet Loss Analysis with respect to Node Speed: Clearly the line graphs in Fig. 28 depicts that the
UDP has highest packet loss ratio as compared to TCP. TCP packet loss ratio is same relatively from 10 to
50 m/s and also the values are nearest to zero. On the other hand, UDP packet loss rate is becoming lower
by more speed. The packet loss of UDP at 10, 20, 40 and 50 m/s is less than the packet loss at 30 m/s.

TCP Pkt loss

60 UDP Pkt loss

Packet loss(Nos of Packets)

50

40

30

20

10

0
30 40 10
50 20
Node Speed
Fig. 28: Packet loss Comparison of DSR over TCP & UDP Connection in case of Node Speed

3.12 DSR Traffic Analysis with respect to Pause Time:

I evaluate Throughput, Delay and Packet Loss with respect to Pause Time.

DSR Throughput Analysis with respect to Pause Time: This is shown in Fig. 29 that comparatively TCP
has more throughputs as compared to UDP. By increasing Pause Time TCP and UDP both degraded as
their throughput become decreases by increases pause time. On some points like Pause Time 250 and 350
seconds the throughput of TCP and UDP is about same.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 68

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP throughput
UDP Throughput

700
600

Throughput(Kbps)
500
400
300
200
100
0
0 50 100 150 200 250 300 350 400
Pause Time
Fig. 29: Throughput Comparison of DSR over TCP & UDP Connection in case of Pause Time

DSR Delay Analysis with respect to Pause Time: It is cleared from the line graphs of Fig. 30 that TCP
has best performance while UDP has degraded performance. TCP and UDP both are affected by the Pause
Time as both increases by increasing Pause Time. Individually, this effect of Pause Time in DSR over TCP
is little while in DSR over UDP it is much. Simply we can say that TCP is little affected from pause time
and UDP is more affected from pause time. So in high Pause Time TCP should be adopted for best
performance.
TCP Delay
UDP Delay
12

10

8
Delay(sec)

0
0 50 100 150 200 250 300 350 400
Pause Time
Fig. 30: Delay Comparison of DSR over TCP & UDP Connection in case of Pause Time

DSR Packet Loss Analysis with respect to Pause Time: Fig. 31 depicts Packet loss Vs Pause Time. It is
clear from the line graphs that about on all given Pause Time the packet loss ratio of TCP is same and
approximately equal to zero. On other hand the Packet Loss ratio of UDP is very higher than that of TCP. It
is also cleared from the Figure that UDP performance become well (good) by increasing Pause Time.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 69

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

TCP Pkt loss

UDP Pkt loss
60

Packet loss(Nos of Packets)

50

40

30

20

10

0
0
50 100 150 200 250 300 350 400
Pause Time
Fig. 31: Packet loss Comparison of DSR over TCP & UDP Connection in case of Pause Time

Conclusion
The analysis of Ad hoc routing protocols is done in the realistic scenario of WSN. After doing the
simulation based analysis of AODV, DSR and DSDV in realistic scenario of WSN. Packet delivery Ratio
of AODV is better than other three protocols so we can say this protocol is applicable to carry sensitive
information in WSN but it fails for the scenario where transmission time should be very less as it has
highest end to end delay. For quick transmission DSR performs well but not suitable to carry information
as packet loss is very high. We can see that TCP over DSDV should be adopted for high throughput. Node
density has higher effect on the delay of UDP as compared to TCP so in high delay, UDP should not be
used. The simulation results show that in high mobility scenarios TCP has high throughput than UDP so in
such situation TCP is recommended by our simulation setup. Also in high mobility environment the
performance of TCP is better than UDP. Pause Time has minor effect on the throughput, delay and packet
loss of the TCP as compared to UDP.

References
Bijan Paul, Md Ibrahim, Md. Abu Naser Bikas, “Experimental Analysis of AODV & DSR over TCP &
CBR connections with varying speed and node density
Chadha, Manveen Singh, and Rambir Joon. "Sandeep.,“Simulation and Comparison of AODV, DSR and
AOMDV Routing Protocols in MANETs”." International Journal of Soft Computing and Engineering,
vol.2, no.3,pp. , July 2012
Gupta, Sachin Kumar, and R. K. Saket. "Performance Metric Comparison of AODV and DSDV routing
Protocols in MANETs Using ns-2," IJRRAS, Vol.7, no. 3, pp.339-350, 2011.
Haas, Z.J.; Halpern, J.Y.; Li, L.” Gossip-Based Ad Hoc Routing”. In Proceedings of the 19th Conference of
the IEEE Communications Society (INFOCOM), New York, NY, USA, June 2002; 23–27 ,pp. 1707–
1716.
H.Nakayama, N. Ansari, A. Jamalipour, Y. Nemoto, N. Kato, “On data gathering and security in wireless
sensor networks”, in :Sarnoff symposium, IEEE, 2007.
Intanagonwiwat, C.; Govindan, R.; Estrin, D.; Heidemann, J. Directed diffusion for wireless sensor
networking. IEEE/ACM Trans. Netw. 2003, 11, 2–16.
J. Liu and I. Chlamtac, Mobile Ad-Hoc Networking with a View of 4G Wireless: Imperatives and
Challenges, Mobile Ad Hoc Networking, Wiley-IEEE Press, July 2004

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 70

I
R
B
A www.irbas.academyirmbr.com June 2015

S International Review of Basic and Applied Sciences Vol. 3 Issue.6

K. Khamforoosh, and H. Khamforoush, “A new rounting Algorithm for Energy Reduction in Wireless
Sensor Networks”, IEEE, 2009
Kulkarni, ShrirangAmbaji, and G. Raghavendra Rao., "A Performance Analysis of Energy Efficient
Routing in Mobile Ad hoc Networks," International Journal of Simulations–Systems, Science and
Technology (IJSSST), vol. 10, no.1, pp.1-9, 2010.
Kumar, Anit, and Pardeep Mittal., "Evaluating AODV & DSR Routing Protocols in Mobile Ad-Hoc
Networks Using TCP Traffic Class,"International Journal of Advanced Research in Computer Science
and Software Engineering, vol.3, no. 5, pp. 66-70, June 2013.
Kumar, Ajay, and Ashwani Kumar Singla., "Performance Evaluation of MANET routing Protocols on the
Basis of TCP Traffic pattern," International Journal of Information Technology Convergence and
Services (IJITCS), vol.1, no. 5, pp.41-48, October 2011.
K. R. Chowdhury and I. F. Akyildiz, "Interferer Classification, Channel Selection and Transmission
Adaptation for Wireless Sensor Networks," in IEEE ICC ’09, Dresden, Germany, 2009.
Kulik, J.; Heinzelman, W.R.; Balakrishnan, H.” Negotiation based protocols for disseminating
information in wireless sensor networks”. Wirel. Netw. 2002, 8, 169–185.
Sun, Dong, and Hong Man. "TCP flow-based performance analysis of two on-demand routing protocols for
mobile ad hoc networks." Vehicular Technology Conference, 2001. VTC 2001 Fall. IEEE VTS 54th.
Vol. 1. IEEE, 2001.
Vijaya, Amiya kumar Rath, Pinak Bhusan Mishra, Amulya Ratna Dash (2011) 2nd international conference
on Emerging Applications of Information technology, IEEE
V.K.Taksande & Dr. K. D. Kulat “Performance Comparison of DSDV, DSR, AODV Protocol with IEEE
802.11 MAC for Chain Topology for Mobile Ad-hoc network using NS-2 ” IJCA Special Issue on
“2nd National Conference- Computing, Communication and Sensor Network”CCSN, 2011.
Zahid, Haleem Farman, et al. "Performance Evaluation of TCP (Transmission Control Protocol) and UDP
(User Datagram Protocol) over Destination Sequence Distance Vector (DSDV) for Random Waypoint
Mobility Model." World Applied Sciences Journal, vol.20, no. 7 , pp. 910-916, November 2012.
Zeenat Rehena, Sarbani Roy, Nandini Mukherjee ,”A Modified SPIN for Wireless Sensor Networks”, 978-
1-4244 8953-4/11 2011 IEEE.

ISSN: 2308-7056 Abbas, Ahmad, Bhatti & Akhtar (2015) 71